Tektronix 2465B oscilloscope teardown

[med6753]

And now the final part: cooling of the U800.
The consensus seem to be, that the TekTronix made chips are "better" in terms of failing on the long term.
Mine is not getting hot, but it may have to do whem touching the chip is is just doing almost nothing?
Maybe when the scope is running on a very high frequency with a high repetition rate, it may get hot due to more capacitive load??

Anyway, I decided to help this little chip a bit.
What I did was, I removed the nut where the metal tab from under the chip is, and made a brass (for thermal conductivity) bushing the same height as the remaining part of the chip.
So basically, I try to guide the heat through this bushing to the top, and there transfer it to a cooling element.
I also put some thermal conductive grease on top of the chip, although this may be of little effect.
The cooling I found at Mouser's, it is a double to-220 part.

I will start adding pictures to show what I did.

Un saudo,

Satbeginner (Leo)

Be careful. The studs that secure U800 are NOT at ground potential. I may be mistaken but I think each stud is at a different potential. So you could easily cause a short. The heatsink modifications I've seen for U800 are glued to the top of the chip away from the studs. So check what you've done very carefully or you might have magic smoke.   

[med6753]

I thought the Maxim dies were not specific to the U800 but for the other custom chips, U400's maybe

I've read that when Tek stopped manufacture of their own custom dies Maxim took over and that includes U800. If I'm wrong someone correct me.   

[BravoV]

and funny enough, one of the 15 Ohm resistors went up to 4968 Ohm!! without any markings or change of color.

Leo, which resistor is that ?

No markings ? You still have it ? Photo please.

[BravoV]

Be careful. The studs that secure U800 are NOT at ground potential. I may be mistaken but I think each stud is at a different potential. So you could easily cause a short. The heatsink modifications I've seen for U800 are glued to the top of the chip away from the studs. So check what you've done very carefully or you might have magic smoke.

That studs (or embedded bolts ?) is -5 (negative five) volts.

Posted this at -> post #2 at the beginning of this thread, yep, I measured it with DMM.

[Satbeginner]

Be careful. The studs that secure U800 are NOT at ground potential. I may be mistaken but I think each stud is at a different potential. So you could easily cause a short. The heatsink modifications I've seen for U800 are glued to the top of the chip away from the studs. So check what you've done very carefully or you might have magic smoke.

That studs (or embedded bolts ?) is -5 (negative five) volts.

Posted this at -> post #2 at the beginning of this thread, yep, I measured it with DMM.

Sorry, my bad, wrong post!!


Just one of the 1/4W resistors used at the line input filter

[Satbeginner]

And now the final part: cooling of the U800.
The consensus seem to be, that the TekTronix made chips are "better" in terms of failing on the long term.
Mine is not getting hot, but it may have to do whem touching the chip is is just doing almost nothing?
Maybe when the scope is running on a very high frequency with a high repetition rate, it may get hot due to more capacitive load??

Anyway, I decided to help this little chip a bit.
What I did was, I removed the nut where the metal tab from under the chip is, and made a brass (for thermal conductivity) bushing the same height as the remaining part of the chip.
So basically, I try to guide the heat through this bushing to the top, and there transfer it to a cooling element.
I also put some thermal conductive grease on top of the chip, although this may be of little effect.
The cooling I found at Mouser's, it is a double to-220 part.

I will start adding pictures to show what I did.

Un saludo,

Satbeginner (Leo)

Be careful. The studs that secure U800 are NOT at ground potential. I may be mistaken but I think each stud is at a different potential. So you could easily cause a short. The heatsink modifications I've seen for U800 are glued to the top of the chip away from the studs. So check what you've done very carefully or you might have magic smoke.

Hi,

thanks for your response, in the picture you can see this cooler stays far away from any surrounding part, I also have like 8-9 mm from the case.
Although my U800 stayed cool, I just wanted to help it a little by conducting heat away from the chip.
The spring washers are no good in doing this, but -my thoughts- a solid brass ring might help, and also cooling the chip -although plastic- itself a bit.

I never liked glue and heatsinks, in my book they always end up somewhere loose in the device... ;-)

Un saludo,

satbeginner (Leo)

[Satbeginner]

and funny enough, one of the 15 Ohm resistors went up to 4968 Ohm!! without any markings or change of color.

Leo, which resistor is that ?

No markings ? You still have it ? Photo please.

Hi,

it was just one of the 1/4W resistors used in the line input filter.

Un saludo,
Satbeginner (Leo)

[cheeseit]

Must be either R1010 or R1019. Both were bad in one of my 2465B, one was open and the other measured 68K, but both had visible signs although not that easy to see.

[tggzzz]

Must be either R1010 or R1019. Both were bad in one of my 2465B, one was open and the other measured 68K, but both had visible signs although not that easy to see.

The problem in my 2465 was R1016. It was easy to determine, since it was missing. So was a circular patch of PCB pre-preg, and C1016 was not a cuboid.

Tektronix designers thoughtfully provided a big pointer to this, in the form of an irregular large black marking on the PCB and a smoke signal that was visible from outside the case.

[Satbeginner]

Must be either R1010 or R1019. Both were bad in one of my 2465B, one was open and the other measured 68K, but both had visible signs although not that easy to see.

It was R1010, now this one and R1019 both replaced with 2W ceramic ones.

[Satbeginner]

And now the final part: cooling of the U800.
The consensus seem to be, that the TekTronix made chips are "better" in terms of failing on the long term.
Mine is not getting hot, but it may have to do whem touching the chip is is just doing almost nothing?
Maybe when the scope is running on a very high frequency with a high repetition rate, it may get hot due to more capacitive load??

Anyway, I decided to help this little chip a bit.
What I did was, I removed the nut where the metal tab from under the chip is, and made a brass (for thermal conductivity) bushing the same height as the remaining part of the chip.
So basically, I try to guide the heat through this bushing to the top, and there transfer it to a cooling element.
I also put some thermal conductive grease on top of the chip, although this may be of little effect.
The cooling I found at Mouser's, it is a double to-220 part.

I will start adding pictures to show what I did.

Un saludo,

Satbeginner (Leo)

Be careful. The studs that secure U800 are NOT at ground potential. I may be mistaken but I think each stud is at a different potential. So you could easily cause a short. The heatsink modifications I've seen for U800 are glued to the top of the chip away from the studs. So check what you've done very carefully or you might have magic smoke.

Hi,

thanks for your response, in the picture you can see this cooler stays far away from any surrounding part, I also have like 8-9 mm from the case.
Although my U800 stayed cool, I just wanted to help it a little by conducting heat away from the chip.
The spring washers are no good in doing this, but -my thoughts- a solid brass ring might help, and also cooling the chip -although plastic- itself a bit.

I never liked glue and heatsinks, in my book they always end up somewhere loose in the device... ;-)

Un saludo,

satbeginner (Leo)

I found this message thread in the Tekscopes Yahoo group.
Does make you think...

https://groups.yahoo.com/neo/groups/TekScopes/conversations/topics/120573

It appears there are 2 main causes for a U800 to die:
- heat, and as a result of that the internal separation of the chip and the substrate (therefor also torque is very important!), and
- internal arcing of the tube...

Well, I hope if the scope and the tube are old, the tube will be properly conditioned so the arcing will be gone by now.

In that thread I also found some people who did the same kind of cooling as I did.

Un saludo,

Satbeginner (Leo)

 

[med6753]

You can't view that link to the Tek Yahoo group unless you are a member. I was a member but for reasons that I will keep to myself I left the group. With one exception I found the blog to be more helpful in fixing issues with my 2465. And perhaps that's due to most of my issues were with the DMS version that has Counter/Trigger and DMM options which I'm finding are not that common.

U800: Opinions range from there is no issue to the sky is falling. My take....depends if you are dealing with Tek die or Maxim die and you must be proactive in keeping the cooling system serviced. Make sure all vent holes are clear and the fan is operating properly. I tend to fall into the position that if it ain't broke, leave it alone. In my case it's been there since 1985 and works fine. Would an additional heatsink help? I'm sure it would but there are risks. Once risk I see is accidentally applying an uneven torque to the mounting studs. And as far as I know the service manual does not list a torque spec for those studs. Too risky IMHO especially if the prime failure seems to be die separation. But I certainly won't fault anyone who want's to go this route.   

[peter-h]

I have open up mine at last. Some chips are date coded 1997. The battery backed up SRAM is 1995.

I believe this scope was "overhauled" in some way about 10 years ago, by the refurbish company which I bought it from.

Some pics here - the site crashes when I try to upload anything sizeable.

http://peter-ftp.co.uk/screenshots/2017-03-28_225205.jpg
http://peter-ftp.co.uk/screenshots/2017-03-28_225239.jpg

[tautech]

I have open up mine at last. Some chips are date coded 1997. The battery backed up SRAM is 1995.

I believe this scope was "overhauled" in some way about 10 years ago, by the refurbish company which I bought it from.

Some pics here - the site crashes when I try to upload anything sizeable.

http://peter-ftp.co.uk/screenshots/2017-03-28_225205.jpg
http://peter-ftp.co.uk/screenshots/2017-03-28_225239.jpg

You're constrained by the forum 1Mb file size limit, compress the size of your pics to under 1 Mb and much less unless great detail is needed. In most cases 100Kb is plenty big enough.

[BravoV]

I have open up mine at last. Some chips are date coded 1997. The battery backed up SRAM is 1995.

I believe this scope was "overhauled" in some way about 10 years ago, by the refurbish company which I bought it from.

Some pics here - the site crashes when I try to upload anything sizeable.

http://peter-ftp.co.uk/screenshots/2017-03-28_225205.jpg
http://peter-ftp.co.uk/screenshots/2017-03-28_225239.jpg

Interesting, your Maxim U800 date code is 95 ? Manufactured in 1995 ?

Anyone know approx. when Tek stopped producing 2465B/2467B scopes ?

[peter-h]

"There are 6 cascaded 74F164 shift registers that drive the LEDs directly and are loaded by LED CLK and LED DATA."

These are not on the CPU board and are probably quite hard to get to. I have dismantled the "easy" covers...

The capacitors look OK
http://peter-ftp.co.uk/screenshots/2017-03-29_073937.jpg
http://peter-ftp.co.uk/screenshots/2017-03-29_074006.jpg

The scope shows 1995 all over the place
http://peter-ftp.co.uk/screenshots/2017-03-29_074103.jpg
but there are some 1997 chips in there. So maybe it was built up from modules sitting on the shelf.

Now I need to find out where to check the supply voltages, and how to get to the shift registers that drive the display. It appears that to get access to the front panel PCB one has to peel off the front label (there seem to be screws under it) which is going to be awfully messy.

Thanks for the 1MB image size input. For now I am using screenshots on my FTP site, which should not go missing because I use it for various other stuff.

[BravoV]

Now I need to find out where to check the supply voltages,

The J119 for all voltage rails.

Don't forget to get a big fan to blow the whole A1 board while you're working on it, remember those poor hybrids.

[peter-h]

Super many thanks!
All are in spec except the +10V which is 9.987V - verified with two meters one of which is a HP3468A which has been checked against a 0.01% reference.

http://peter-ftp.co.uk/screenshots/2017-03-29_084017.jpg
The missing minus sign on the last line of that pic is my mistake



The hybrid chips didn't get particularly warm - I would estimate +45C. And the big plastic one with the two screws was not warm at all.

The 74F164 chips are indeed on the front panel PCB - A6



Now I am trying to find the clock and data signals going to the front panel, without having the dismantle the front panel which is obviously complicated.

Would it be around here?



Obviously I am hoping the fault is with the clock and data rather than after the shift registers.



I found the LED clock on the CPU board as pin 16 J652



but can't find the LED DATA signal.

Anyway, U2550 (HC138) has +5V on pin 14 which is the LED CLOCK, so definitely there is no clock signal to the LEDs. Other outputs are waggling OK. I will try changing this chip.

EDIT: I found there is indeed no clock unless the front panel is operated. Then I do see pulses, so the clock is OK. I just need to find where the LED data is... The troubleshooting flow chart refers to P652 pin 13 and pin 5 (for LED clock and data) but thise don't correspond to the circuit. Also the IDC connectors on the CPU board can't be probed; one has to find the signals on a chip.

[MarkL]

...
but can't find the LED DATA signal.

When you say you "can't find" it, do you mean there's no signal or you can't locate it on the schematic?  LED DATA is on the snippet you posted; outlined below.

When checking signals, make sure they are true TTL, low < 0.7V and high > 2.1V.  If you're having a power issue or a chip somewhere without power, it can lead to strange logic levels.  In this case, LED CLOCK and DATA are coming from HCT series chips (CMOS TTL) so they should be very near 0V for low and 5V for high.

Maybe you could share some screen captures?

You can see in the snippet that CONT DATA (control data) signal is the same as LED DATA.  CONT DATA, among other things, is used to control the attenuator relays.  Can you hear the relays clicking if you cycle through the V/div settings?  That would be a good indication CONT DATA, and therefore LED DATA, is ok as far as U2210.

I would also check for continuity from GND to J652.19, and from +5V to J652.20 to make sure the front panel is getting power (or at least getting to J652).  Use the power connector where you measured the incoming voltages as your reference points.

As you're poking around, keep in mind a lot of failures in these circa scopes is caused by bad PC traces.  Don't assume that if a signal or power is ok in one place, then it's ok everywhere.

[peter-h]

Mark - many thanks.

"When you say you "can't find" it, do you mean there's no signal or you can't locate it on the schematic?  LED DATA is on the snippet you posted; outlined below."

I did find the likely LED DATA signal later Pin 5 J652. I need to probe it on pin 2 U2210.

Is the LED clock definitely pin 16 J652?

"Maybe you could share some screen captures?"

Stand by - in a few hours.

"You can see in the snippet that CONT DATA (control data) signal is the same as LED DATA.  CONT DATA, among other things, is used to control the attenuator relays.  Can you hear the relays clicking if you cycle through the V/div settings?  That would be a good indication CONT DATA, and therefore LED DATA, is ok as far as U2210."

Yes the buttons seem to work for clicking various relays e.g. the input AC or DC or GND mode. So it has to be the clock - or something actually on the front panel.

"I would also check for continuity from GND to J652.19, and from +5V to J652.20 to make sure the front panel is getting power (or at least getting to J652).  Use the power connector where you measured the incoming voltages as your reference points."

How do you get to the front panel? I can see some screws on the bottom, but there are many buttons etc to remove. At that point I would rather pay someone who has done this before, a couple of hundred GBP... I am in the UK.

"As you're poking around, keep in mind a lot of failures in these circa scopes is caused by bad PC traces.  Don't assume that if a signal or power is ok in one place, then it's ok everywhere."

Sure... OK. The internals all seem very clean on this one.

[MarkL]

Is the LED clock definitely pin 16 J652?

I think so.  I would believe the schematic before the troubleshooting chart.  The J652.13 referenced in the chart is already spoken for as ROW 0 at the bottom right corner of schematic <2>.

You could unplug J652 and do a quick continuity check from J652.16 to U2550.14 to confirm.

How do you get to the front panel? I can see some screws on the bottom, but there are many buttons etc to remove. At that point I would rather pay someone who has done this before, a couple of hundred GBP... I am in the UK.

There are step-by-step instructions starting on page 6-28 in the maintenance section.  I've never had the privilege of needing to remove the front panel.  If there are any tricks to it perhaps someone who has done it can post.

[peter-h]

I do get 0V to +5V transitions on the clock and data.

Regarding the continuity check, where is pin 1 of J652? I assumed pin 1 is the coloured conductor on the ribbon cable. There is continuity OK.

So the LED signals seems OK, which is bad news.

I am pretty sure what I am seeing on the LEDs is that they are all dimly lit but that is intentional - it is a form of backlighting. Then, the selected mode is illuminated brightly, and that is what is not happening.

I have been trying to scope it with a TDS2004B which is a really horrible digital scope and it is really difficult to get it to show any low duty cycle signals. I just don't know how to use it. It was used purely as an FFT display on a vibration analyser. So I am using the 2465B to probe itself and that works. All the front panel switches and modes seem to work, although of course one cannot see the one selected; one has to go to the top of a column with lots of presses and then e.g. (on the trig mode) press 3 times to get NORM

So the front panel will need dismantling... I am not sure I have the enthusiasm for that.

[james_s]

It sounds like some noise is getting into the shift register clocking to the LEDs. You should be able to capture a run of the clock and data lines with the DSO and see pretty clearly which LEDs should be on, the data is shifted out as a chunk of 48 bits, so you should clearly see groups of 48 clock pulses with the data line transitioning high or low for each LED lit. If you force the LED data line either low or high, all LEDs should either light full brightness or go out. I don't think the dim glow is normal, none of my Tek scopes do that.

[Satbeginner]

I recently removed my front panel to replace one of the plastic extension shafts.

I had to remove the surrounding plastic front trim by removing the screws all around the scope.

After that I removed the two flatcables who plug into the A5 board.

Now you can slide out the entire control unit.

That can be divided into the two boards by removing the 4 screws.
Watch out you do not damage the fragile flatcables.

Succes,

Satbeginner

[peter-h]

So... one should be able to access the front panel electronics without having to remove all the buttons? In that case I will give it a go.

There are definitely not 48 pulses - unless there are long gaps between some of them. I will try to capture it.